More False-Positive Problems
SIR—Sarkar and Sommer have reported in Scientific Correspondence［1］the use of ultraviolet irradiation as a sterilization technique for the control of contamination in the polymerase chain reaction（PCR），but their procedure does not deal with all the problems.First，they introduce true target DNA and Taq DNA polymerase after irradiation，which will produce the same level of sporadic false-positive PCR signals even though all the contaminating carryover molecules in the other PCR reagents are sterilized.Second，their conclusion that the oligonucleotide primers for the PCR retain their full functional integrity after irradiation is not warranted.Primer damage which leads to a compromise in signal sensitivity can be evaluated only when the PCR amplification is limited so that the concentration of product is maintained well below PCR plateau concentrations.Finally，the critical nature of the size and sequence specificity of the PCR product being inactivated was not addressed.Here，we elaborate only on the last point（see Figure 1）.
Figure 1 Effect of PCR product length on UV sterilization of PCR products.Two PCR products were examined：a 115 mer（Ou et al.［6］）and a 500 mer（kit N801-0055，Perkin-Elmer Cetus）.PCR templates were irradiated for 0 min（lanes 1 and 5）；5 min（lanes 2 and 6）；20 min（lanes 3 and 7）；30 min（lanes 4 and 8），amplified for 30 cycles and analysed by PAGE and autoradiography.In panels a（115 mer template），b（plasmid template）and c（λDNA template）.Lanes 1-4 and 5-8 correspond to 105 and 103 copies of PCR template molecules，respectively.Detection sensitivity with our PCR procedure is at least 100 copies.Experimental details available from authors upon request
Pyrimidine dimers undoubtedly contribute to ultraviolet-induced sterilization by functioning as termination sites during the extension reactions of the PCR procedure.Only a fraction of all pyrimidines within a DNA strand form dimers and，as dimers are both made and broken by ultraviolet exposure，this fraction establishes a steady-state level which varies with the irradiation wavelength，the type of pyrimidine dimer and the nucleotide sequences next to the dimer site［2，3］.At steady-state，an upper limit for the number of all dimer defects in a long，irradiated DNA molecule is less than 0.065 defects per base pair［4］.Nondimer photodamage（for example non-cyclobutane-type pyrimidine adducts，thymine glycols，interstrand and intrastrand DNA-DNA crosslinks and DNA strand breaks）can also be termination sites for Taq DNA polymerase.The number of these sites should be at least equivalent to the number of dimer sites［5］.
此文发表于1990年“Nature”345：773-774.
If these defects（0.13 total defects per base pair）are randomly distributed throughout a DNA molecule，a 500-base-long oligonucleotide would have an average of 32 damaged sites.In a population of 105 DNA strands containing an average of 32 termination sites per strand，there is a minute probability that any strand is without at least one termination site.By comparison，100-base-long PCR products will have an average of only 6 modified sites per strand at this same defect density.In a population of 105 molecules with just six average termination sites per strand，a statistically large number of molecules will contain no modified sites and therefore are not sterilized.(https://www.daowen.com)
Furthermore，to be effective for PCR sterilization，the photoinduced defects should be in the sequence region bounded by the 3'ends of the PCR primers.These considerations predict that ultraviolet irradiation would be effective for long but not for short PCR products.Our experiments demonstrate that this is the case.
We irradiated targets with either 254，300 or a combination of 254 and 300-nm bulbs.After irradiation，the remaining PCR reaction components were added to each tube and amplification was performed.As shown in the Figure 1，a 115 mer PCR product is efficiently generated from as little as 103 target molecules that had been irradiated with a combined 254 and 300 nm exposure independent of whether the target is a synthetic 115 mer or an 11-kb plasmid.Irradiation of linerizedλ DNA from which a 500 mer PCR product is made results in 103 copies of theλ being sterilized with the combined 254/300 nm procedure.But 105 copies of theλ-DNA target show an exposure dependent sterilization.Only 30 minutes of exposure is sufficient to reduce the PCR signal below the sensitivity level of the gel assay.
These data imply that caution is needed when applying direct ultraviolet irradiation as a sterilization process in a preamplification mode.Both the length of the PCR product and its internal sequence must be considered.The 115 mer we used was chosen because it has a very high fraction of adjacent thymines between the prime regions of the PCR product.But the distribution of thymines is such that most of the potential pyrimidine dimers will occur on one strand of the product.This type of distribution must also be taken into account when considering ultraviolet sterilization techniques.
